The postsynaptic density (PSD) at excitatory glutamatergic synapses is a large molecular machine with a mass greater than one billion Daltons. The PSD is known to be a key site of information processing and storage. In order to explore the detailed molecular organization of the PSD, we developed a method to freeze-substitute hippocampal cultures and then examine them in thin sections by EM tomography to show individual protein complexes in their natural setting within the PSD. The initial work employing tomography revealed that the core of the PSD is an array of vertically oriented filaments that contain the scaffold protein, PSD-95, in an extended configuration and a polarized orientation, with its N-terminus positioned at the postsynaptic membrane. This finding provided insight into the overall organization of the PSD because scaffolding proteins such as PSD-95 family MAGUK proteins have distinct multiple, diverse binding sites for other proteins arrayed along their length. Thus, the regular arrays of PSD-95, perhaps with other family members, impose an ordering on many other PSD proteins, including the glutamate receptors, and provide an overall plan for the structure of the PSD. FRET constructs were made to study possible mechanisms that regulate PSD-95 MAGUK conformations in collaboration with the William Green. Two fluorophores (RFP and YFP) are fused to the opposite termini of PSD-95 or other family members. The labels allow the conformations of the family members to be determined both by immunogold-EM and by making FRET measurements on living cells (if the two ends of the molecule FRET, they must be in a closed configuration). So far, results suggest that PSD-95 adopts an extended conformation in PSDs, but in closed conformation at non-synaptic sites. In contrast, SAP-97, another MAGUK has an open configuration, but is oriented parallel with the post synaptic membrane. Open conformation of PSD-95 at the PSD is a requirement for it to interact with NMDAR and AMPAR-Stargazin complexes. EM tomography also revealed that the C-terminal ends of the PSD-95 vertical filaments are associated with horizontally oriented filaments. One class of horizontal filament is ordered to form hexagonal cross-linkers with the vertical filaments, and is concentrated beneath the NMDA receptors. Immunogold labeling identifies a class of horizontal filaments as GKAP, which is a known to bind to the GK domain at the C-terminal end of PSD-95. Immunogold labeling is also being used to identify Shanks and GKAP in tomographic reconstructions. The emerging structural model of the PSD shows how the PSD-95 matrix can stabilize glutamate receptors, and at the same time allows room for the addition of new receptors at the edges of the PSD. Identification of the components of the PSD is time consuming and the methods for identifying the proteins need improvement. An expressible prob, miniSOG, confirms that the vertical filaments are PSD-95. We are now preparing probes to use miniSOG to definitely identify GKAP and SHANK in the PSD. The idea that the PSD-95 dependent scaffold stabilizes the PSD has been explored by using EM tomography to determine the effects of RNAi knock down of MAGUKs. Recently, we examined the effects of knocking down simultaneously three major MAGUK proteins: PSD-95, PSD-93 and SAP102, and for the first time, EM tomography revealed significant loss from the central core of the PSD, including NMDA receptor structures, vertical filaments, and AMPA receptors. Electrophysiology measurements by collaborators from the Nicoll laboratory (UCSF) characterizing the effects of the same knock down show significant functional loss of NMDAR and AMAPR type EPSPs at levels compatible with the structural losses. Electron microscopy, showed depletion of vertical filaments along with AMPAR type structures at the peripheral region of the PSD and significant reduction of size of NMDAR clusterin the middle of the PSD. These structural data indicate that vertical filaments corresponding to MAGUKs anchor AMPARs and are also a factor in organizing NMDARs. Thus, PSD-95 MAGUKs are demonstrated to be the essential organizer of glutamate receptors at the PSD. A new electron microscopic method in collaboration with Richard Leapman using high voltage STEM tomography (HVST) is compatible with sections up to two micrometers thick and is revealing detailed reconstructions of whole synapses. We used HVST on 1-2 um sections that contain entire PSDs at synapses to demonstrate that simultaneous knock down of the three major PSD-95 family MAGUKs results in significant reduction in the overall PSD area, leaving many synapses with only small PSDs. Since a subpopulation of synapses completely loses their electron dense PSD material, the triple knock down in effect results in many silent synapses as reported by electrophysiology. In collaboration with the National Institute of Standards and Technology quantitative mass spectrometry was applied to evaluate stoichiometries of PSD components. Copy numbers for targeted proteins within the PSD were subsequently estimated using a consensus literature value for the copy number of PSD-95. The NMDA receptor to AMPA receptor copy number ratio was determined to be 1:2, yielding an estimate of 34 10 NMDA channels and 68 36 AMPA channels per average PSD, respectively, in line with estimates from other methods. A ratio for AMPAR tetramers to TARP auxiliary subunits was 1:2 supporting the assertion that most AMPA receptors anchor to the PSD via TARP subunits. The study also generated estimates for the for the copy number of several key PSD proteins for the first time, and confirmed or challenged, previous estimates for certain other PSD components. These copy numbers provide valuable guidelines to map quantitatively molecular architecture of the PSD. A new initiative focuses the on molecular organization at the synaptic cleft. A preliminary study by EM tomography classified structures spanning the cleft of excitatory synapses is now published (High et al 2015). A more elaborate molecular characterization of cleft architecture requires knowledge of the stoichiometry of candidate proteins at the cleft. To this end, a quantitative mass spectrometric approach will be applied, similar to the one we used to estimate average copy numbers of proteins at the PSD (Lowenthal et al 2015). As a first step, a new method for the isolation of a synaptic junction fraction is being developed, because existing strategies rely on the use of detergents that cause damage to the cleft. Use of phospholipase A2 instead of detergent to selectively remove peripheral membranes, while keeping junctional membranes intact, appears to give optimal results. Current work is directed to characterize this synaptic junction fraction through biochemical and EM analyses. A second new initiative is being explored in collaboration with Carolyn Smith in the NINDS Light Mocroscopy Facilty. Dr Smith has cultured a primitive animal that lacks synapses but shows behavior suggestive of neural function. Thus, it would appear that there is an initial step in evolving nervous systems, prior to the development of synapses. that appears to involve peptide signaling. Knowing exactly how these sysstems function may provide information on peptide signaling in mammalian brains that has been overlooked.